A New Look at Quasars
Recent observations from the Hubble Space Telescope
may reveal the nature and origin of quasars,
the mysterious powerhouses of the cosmos
by Michael Disney
52
Scientific American
June 1998
Copyright 1998 Scientific American, Inc.
JOHN BAHCALL Institute for Advanced Study AND NASA
DON DIXON
Q
uasars are the most luminous objects in
the universe. They give off hundreds of
times as much radiation as a giant galaxy like our own Milky Way, which is
itself as luminous as 10 billion suns. Nevertheless,
by astrophysical standards, quasars are minute objects, no more than a few light-days in diameter, as
compared with the tens of thousands of light-years
across a typical galaxy. How in heaven can they
generate so much energy in such tiny volumes?
What are they, and can they be explained by the
ordinary laws of physics? To answer these questions, astronomers are training their most advanced instruments—the Hubble Space Telescope
in particular—on these celestial superstars.
The first quasar was discovered in 1962, when
Cyril Hazard, a young astronomer at the University of Sydney, began to study a powerful source of
radio waves in the Virgo constellation. Hazard
could not pinpoint the source, because the radio
telescopes of the time were not precise enough, but
he realized that the moon would occult the unknown object when it passed through Virgo. So he
and John Bolton, the director of a newly built radio telescope in Parkes, Australia, pointed the instrument’s giant dish toward the radio source and
waited for the moon to block it out. By timing the
disappearance and reappearance of the signal,
they would be able to pinpoint the source of radio
emissions and identify it with a visible object in the
sky. Unfortunately, by the time the moon arrived
the great dish was tipped so far over that it was
running into its safety stops. Apparently unperturbed by the risk, Bolton sheared off the stops so
that the telescope could follow the occultation
downward until the rim of the dish almost touched
the ground.
His daring was to be rewarded. From their measurements Hazard was able to calculate the first
accurate position for such a cosmic radio source
and then identify it with a comparatively bright,
starlike object in the night sky. The position of
that object—dubbed 3C273—was sent to Maarten
Schmidt, an astronomer at the Mount Palomar
Observatory in California, who had the honor of
taking its optical spectrum. After some initial puzzlement, Schmidt realized he was looking at the
GALACTIC COLLISIONS may sometimes result in the birth
of a quasar. A massive black hole at the core of one of the
galaxies sucks in stars and gas from the other galaxy, and
the maelstrom of infalling matter generates a beam of intense radiation. Such a process may be occurring in quasar
PG 1012+008 (inset), as observed by the Hubble Space Telescope. The quasar is 1.6 billion light-years from the earth.
Copyright 1998 Scientific American, Inc.
Scientific American
June 1998
53
1013
QUASAR
GIANT
ELLIPTICAL
GALAXY
1012
1011
1010
109
108
1010
1015
1020
FREQUENCY (HERTZ)
1025
QUASAR SPECTRUM of 3C273—one of the brightest quasars
and the first to be discovered—is far broader than the spectrum
of a typical giant elliptical galaxy (left). In the optical range, the
quasar is hundreds of times more luminous. Quasars were most
spectrum of hydrogen shifted redward
by the expansion of the universe. The
16 percent redshift meant that 3C273
was about two billion light-years from
the earth. Given the distance and the observed brightness of the object, Schmidt
calculated that it had to be emitting
several hundred times more light than
any galaxy. The first quasistellar radio
source—or quasar—had been discovered.
Spurred by Hazard’s and Schmidt’s
work, astronomers identified many
more quasars in the following years.
Observers discovered that the brightness
of many quasars varied wildly; some
grew 10 times as bright in just a matter
of days. Because no object can turn itself on and off in less time than it takes
for light to travel across it, the astonishing implication was that these highly
luminous objects must be a mere lightweek or so across. Some reputable as-
200
100
0
0
2
4
6
8
10 12 14
COSMIC TIME (BILLIONS OF YEARS)
the quasar astronomer. First, how are
quasars related to galaxies and stars?
Second, how long does each quasar pour
out its enormous energy? In our immediate cosmic neighborhood—within one
billion light-years from the earth—there
is only one quasar for every million galaxies. But that does not necessarily mean
that quasars are much rarer than galaxies; they could be just as common but
have much shorter luminous lifetimes.
This brings us to the third question: Why
were quasars far more numerous in the
past? At a redshift of 200 percent—about
10 billion light-years away—the number of quasars jumps 1,000-fold. In the
early universe, apparently, quasars were
1,000 times more common than they
are today. And last, the most perplexing
question: How do quasars generate their
prodigious energy?
None of these questions can be easily
JOHN BAHCALL Institute for Advanced Study AND NASA
Scientific American
c
MICHAEL DISNEY AND NASA
b
JOHN BAHCALL Institute for Advanced Study AND NASA
54
300
numerous when the universe was two to four billion years old
(right). Today quasars are 1,000 times less common. Quasars
were also rare in the very early history of the universe, but the
exact numbers are uncertain.
tronomers refused to believe that the
enormous distances and luminosities
implied by the redshifts could be so
great. The controversy spilled over to
the popular press, where it attracted a
younger generation of scientists, like
myself, into astronomy.
Since then, astronomers have catalogued thousands of quasars, some with
redshifts as large as 500 percent. They
are not difficult to find, because unlike
stars, and unlike galaxies composed of
stars, they emit radiation of all energies
from gamma rays to radio. Ironically,
the radio emissions by which they were
first discovered turn out to be, in energetic terms, the least significant portion
of their output. For that reason, some
astronomers argue that the name “quasar” should be superseded by QSO, for
quasistellar object.
There are four big questions facing
a
400
LAURIE GRACE
GAMMA RAY
X-RAY
QUASAR DENSITY PER CUBIC GIGAPARSEC
LUMINOSITY (SOLAR UNITS)
OPTICAL
RADIO
ULTRAVIOLET
MICROWAVE INFRARED
1014
A New Look at Quasars
June 1998
Copyright 1998 Scientific American, Inc.
answered. The typical quasar is so far
from the earth that its image on the
largest ground-based optical telescope
would be 100 million times too small to
be resolved. From the outset, one school
of astronomers felt that quasars had to
be sited in galaxies, probably in their
nuclei. They gathered evidence to show
that all the phenomena observed in
quasars were manifested, albeit in a far
weaker form, in the nuclei of about 1
percent of the giant galaxies near the
Milky Way. A whole zoo of active galactic nuclei were revealed, including radio
galaxies, Seyferts, blazars, optically violent variables, superluminal sources and
so on. But astronomers could not tell
whether these objects were separate
classes of galactic nuclei or representations of the same phenomenon viewed
from different angles or at different stages of development. Nor could astronomers explain the exact relation between
the active galactic nuclei and quasars.
Critics of the theory linking the two
types of objects argued that the luminosity of the active nuclei did not even
approach that of quasars. And the sheer
power of quasars is their most distinctive and mysterious characteristic.
A more direct approach was taken by
Jerry Kristian, another astronomer at
Mount Palomar, in 1973. He argued
that if quasars were inside giant host
galaxies, then the images of the closest
quasars should show a fuzzy halo of
light from the stars in the host galaxy. It
would not be an easy observation, because light from the brilliant quasar,
scattered by the earth’s atmosphere,
would swamp the light from the much
fainter host. Nevertheless, Kristian was
able to demonstrate that the lowest redshift quasars did exhibit this faint, fuzzy
halo. His evidence was not very satisfactory, though, because virtually nothing could be discerned about the host
galaxies, not even whether they were elliptical or spiral.
Troubles with Hubble
W
hen the Hubble Space Telescope
was proposed in the mid-1970s,
most quasar observers expected it to
provide the first clear images of host
galaxies, if they really existed. Indeed,
finding host galaxies became one of the
primary objectives of the telescope. We
on the European space telescope team
designed the Hubble’s Faint Object
Camera with quasars very much in
mind. For instance, we built in a highmagnification focus and a coronograph
specially designed to block off the brilliant light from quasars and thus make
the surrounding hosts more visible.
By then, astronomers suspected that
the only way a quasar could produce so
much energy out of such a tiny volume
was if the object contained a massive
black hole at its core. Such a monster
hole, weighing as much as a billion
suns, would suck in all the gas and stars
in its vicinity. Gas would swirl into the
hole at almost the speed of light, generating intense magnetic fields and huge
amounts of radiation. Donald LyndenBell, then an astronomer at the California Institute of Technology, calculated
that a massive black hole could convert
up to 40 percent of the infalling matter’s
rest-mass energy into radiation. Such a
process would be 400 times more effi-
KIM MCLEOD Wellesley College AND NASA
d
cient than the production of thermonuclear energy in stars. For this reason,
massive black holes became the favored
theoretical explanation for quasars. (All
the other plausible models would rapidly evolve into black holes anyway.)
One problem with the model, though,
was explaining how these monsters
could be fed. A black hole of such enormous mass would tend to swallow up
all the nearby stars and gas and then go
out for lack of fuel. To explore this mystery, the European space telescope team
also built a special long-slit spectrograph
into the Faint Object Camera. This instrument was designed to measure the
rotation speed of material in active galactic nuclei and thus weigh the putative black holes at their cores.
After the much delayed launch of
Hubble in 1990, it was soon discovered
that the telescope’s main mirror had
been incorrectly manufactured. The images were such a travesty that quasar
astronomers were devastated. I, for
one, felt that five to 10 of the most productive years of my astronomical life
had been thrown away through unforgivable incompetence. And many others felt likewise. To its credit, however,
the National Aeronautics and Space
Administration had designed Hubble to
be repairable, and astronauts installed
new cameras with corrective optics in
1993. Unfortunately, none of the special instruments in the original cameras
for observing quasars was recoverable.
If we were still going to search for
quasar host galaxies, we would have to
use the new Wide-Field Planetary Camera, which was not designed for the
job. Nevertheless, two teams set out to
try: a European team headed by myself
and an American team led by astronomer John Bahcall of the Institute
for Advanced Study in Princeton, N.J.
Observing quasar hosts with the Hubble’s new camera was akin to looking
into the headlights of an oncoming car
in a snowstorm and trying to identify its
HOST GALAXIES surround most of the quasars
observed by the Hubble Space Telescope. The
spiral galaxy around PG 0052+251 (a) and the
elliptical galaxy around PHL 909 (b) appear to
be undisturbed by collisions. But a galactic crash
seems to be fueling IRAS 04505-2958 (c). A spiral ring torn from one of the galaxies is below the
quasar; the object above it is a foreground star.
Hubble’s new infrared camera observed another
galactic smashup (d). The dots around quasar PG
1613+658 were caused by diffraction; the colliding galaxy is below it and to the left.
A New Look at Quasars
Scientific American
Copyright 1998 Scientific American, Inc.
June 1998
55
M87
HIGH-SPEED
ELECTRON JET
The Remains of a Quasar?
he active nucleus of M87, a giant
elliptical galaxy in the Virgo cluster
(above), may once have been a quasar.
Astronomers trained the Hubble Space
Telescope’s Faint Object Spectrograph
at the core of M87, which emits a jet of
high-speed electrons. Because the light
from one side of the nucleus was blueshifted and the light from the other side
was redshifted (right), astronomers concluded that a disk of hot gas was spinning around the center of the galaxy at
550 kilometers per second (1.2 million
miles per hour). The high velocity indicated the presence of a massive black
hole, which may have powered a quasar
billions of years ago.
—M.D.
FLUX
(ERGS PER SECOND
PER SQUARE CENTIMETER)
T
manufacturer. Astronomers had to take
several shots of each object, subtract the
high beam—the light from the quasar—
and play with the remaining images on
their computers. In most cases, the final
result contained enough detail to make
out a galactic structure. Sadly, Jerry
Kristian, who pioneered this field, was
killed in an ultralight airplane crash in
California just before the Hubble results were published.
What did the space telescope reveal?
Of the 34 quasars observed, about 75
percent showed the faint, fuzzy halo indicating a host galaxy. The remaining
25 percent showed no such halo, but it
is possible that the quasar’s dazzling
beam is blocking the image in those
cases. About half of the host galaxies
were elliptical, and half were spiral. The
56
Scientific American
APPROACHING
RECEDING
1x10–16
0
5,000
5,100
WAVELENGTH (ANGSTROMS)
quasars with the strongest radio signals
were located primarily in elliptical galaxies, but no other patterns were discernible. Most intriguing, about three
quarters of the host galaxies appeared
to be colliding with or swallowing other galaxies.
This finding had already been reported by John Hutchings and his co-workers at Dominion Astrophysical Observatory in Victoria, Canada, who had
used a ground-based telescope with
adaptive optics to observe quasars. But
the Hubble, with its greater resolution,
provided much more vivid evidence of
the galactic interactions. The images suggest that colliding galaxies supply the
fuel for the quasar’s energy production.
Stars and gas shaken loose by the violence of the impact may be funneling
into a massive black hole at the heart of
one of the galaxies. The infalling matter
then generates the intense radiation.
This process would explain the relative numbers of quasars at different stages in the universe’s history. Immediately
after the big bang, there were no galaxies and hence no galactic collisions. Even
if black holes existed then, there was no
mechanism to funnel material toward
them and turn them into quasars. Consequently, few quasars are observed at
very high redshifts—that is, more than
11 billion years ago. But in the following aeons, galaxies began to assemble
and collide, producing the relatively
large number of quasars observed 10
billion light-years from the earth. Finally, the expansion of the universe carried
most galaxies away from one another,
A New Look at Quasars
June 1998
Copyright 1998 Scientific American, Inc.
IMAGES COURTESY OF HOLLAND FORD Johns Hopkins University; NATIONAL
OPTICAL ASTRONOMY OBSERVATORIES AND NASA; LAURIE GRACE (graph)
VIRGO CLUSTER
reducing the number of galactic collisions—and the number of quasars.
Nevertheless, about one quarter of
the host galaxies observed by Hubble—
such as the spiral galaxy surrounding
the quasar PG 0052+251—show no
sign that they are colliding with another galaxy. It is possible that a faint companion galaxy is present in these cases,
but the quasar’s beam is preventing astronomers from seeing it. Or perhaps
there is an alternative mechanism that
can provide enough fuel to transform a
massive black hole into a quasar. What
we do know for certain is that the vast
majority of galactic interactions do not
seem to produce quasars; if they did,
quasars would be far more common
than we observe.
The scarcity of quasars seems to suggest that massive black holes are a rare
phenomenon, absent from most galaxies. But this supposition is contradicted
by recent evidence gathered by a team of
astronomers led by Douglas Richstone
of the University of Michigan. Combining observations from Hubble with
spectroscopic evidence from groundbased telescopes, the team weighed the
nuclei of 27 of the galaxies closest to
the Milky Way. In 11 of the galaxies
Richstone’s group found convincing evidence for the presence of massive dark
bodies, most likely black holes.
Furthermore, some of those massive
black holes may once have been quasars.
In 1994 a group of astronomers led by
Holland Ford of Johns Hopkins University used Hubble to look into the
heart of M87, a giant elliptical galaxy
in the Virgo cluster, about 50 million
light-years from the earth. The active
nucleus of M87 emits a broad spectrum
of radiation, similar to the radiation
produced by a quasar but with only a
thousandth the intensity. The astronomers discovered that the light from
one side of the nucleus was blueshifted
(indicating that the source is speeding
toward the earth), whereas light from
the other side was redshifted (indicating
that the source is speeding away). Ford
concluded that they were observing a
rotating disk of hot gas. What is more,
the disk was spinning so rapidly that it
could be bound together only by a black
hole weighing as much as three billion
suns—the same kind of object that is believed to be the quasar’s power source.
Billions of years ago the nucleus of M87
may well have been a quasar, too.
The Quasar Quest
T
he recent observations have led astronomers to construct a tentative
theory to explain the origin of quasars.
According to the theory, most galaxies
contain massive black holes capable of
generating vast amounts of energy under very special circumstances. The energy production rises dramatically when
gas and stars start falling into the black
holes at an increased rate, typically about
one solar mass a year. This huge infall
occurs most often, but not always, as a
result of galactic collisions or near misses. Quasars were thus far more prevalent in the epoch of high galaxy density,
when the universe was younger and
more crowded than it is now.
What can be said of the individual
lifetimes of these beasts? Not much for
certain. The observed host galaxies
show no evidence that the quasars have
been radiating long enough to damage
them. The hydrogen gas in the host galaxies, for example, has not been substantially ionized, as it might be if quasars were long-lived. The observation
that so many of the host galaxies are interacting—and the fact that such interactions typically last for one galactic rotation period or less—indicates a quasar
lifetime shorter than 100 million years.
And if the existence of massive black
holes in most galaxies implies a past
epoch of quasarlike activity in each case,
then the small number of observed quasars—only one for every 1,000 galaxies
during their most abundant era—suggests a quasar lifetime of 10 million years
or less. If that number is correct, the
quasar phenomenon is but a transient
phase in the 10-billion-year lifetime of a
galaxy. And although the amount of energy generated by each quasar is tremendous, it would account for only
about 10 percent of the galaxy’s lifetime radiant output.
Obviously, more observations are
needed to test the theory. The Hubble
Space Telescope must be trained on a
wider sample of nearby quasars to
search for host galaxies. The existing
samples of nearby quasars are too small
and too narrowly selected for reliable
conclusions to be drawn, and the distant host galaxies are too difficult to
observe with the current instruments.
Astronomers expect to make new
discoveries with the help of two devices
recently installed on Hubble: the Near
Infrared Camera and Multi-Object
Spectrometer (NICMOS), which will
allow scientists to peer into the nuclei
of galaxies obscured by clouds of dust,
and the Space Telescope Imaging Spectrograph (STIS), which has already
demonstrated its usefulness by detecting and weighing a black hole in a nearby galaxy in one fortieth the time it
would have taken previously. In 1999
NASA plans to install the Advanced
Camera, which will contain a high-resolution coronograph of the kind that
was always needed to block the overwhelming quasar light and unmask the
host galaxies.
On the theoretical side, we need to understand how and when massive black
holes formed in the first place. Did they
precede or follow the formation of their
host galaxies? And we would like a convincing physical model to explain exactly how such black holes convert infalling matter into all the varieties of
quasar radiation, from gamma rays to
superluminal radio jets. That may not
be easy. Astronomer Carole Mundell of
Jodrell Bank Observatory in England
once remarked that observing quasars
is like observing the exhaust fumes of a
car from a great distance and then trying to figure out what is going on under
SA
the hood.
The Author
Further Reading
MICHAEL DISNEY is a professor of astronomy at the
University of Wales in Cardiff, U.K. For 20 years he was
a member of the European Space Agency’s Space Telescope Faint Object Camera team. He received his Ph.D.
from University College London in 1968. His other scientific interests include hidden galaxies, bird flight and
the environmental dangers posed by oil supertankers.
Perspectives in Astrophysical Cosmology. Martin J. Rees. Cambridge
University Press, 1995.
Active Galactic Nuclei. Ian Robson. John Wiley, 1996.
An Introduction to Active Galactic Nuclei. Bradley Peterson. Cambridge University Press, 1997.
Information on the Hubble Space Telescope is available at http://www.stsci.
edu on the World Wide Web.
A New Look at Quasars
Scientific American
Copyright 1998 Scientific American, Inc.
June 1998
57